Using Quality by Design to Optimize Tech Transfer of a BFS Process

Nov. 14, 2012
Catalent worked within the QbD framework to help a customer transfer a new blow-fill-seal (BFS) process.

Technology transfer is a difficult and sensitive process.  For a drug service supply partner, transferring an existing drug product from a customer’s facility to its own manufacturing plants offers an opportunity to increase value by improving process efficiencies and equipment design, and to incorporate operational excellence principles into the work.

Using the principles of pharmaceutical Quality by Design (QbD) can help set guidelines for tech transfer, and ensure that opportunities for process improvement are not overlooked when focusing on budget and deadline.

This article will summarize how Catalent worked within the QbD framework to help a customer transfer a new blow-fill-seal (BFS) process.

In 2010, Catalent was approached by a Fortune 500 drug manufacturer to transfer an existing ophthalmic product from a manufacturing site Catalent’s BFS facility in Woodstock, Illinois.  The product was already marketed commercially and the FDA approval for Catalent manufactured product was to be based on the manufacturing process used at the customer’s site. 

Once the site was selected, the technical transfer process began with a detailed comparison of the client’s current manufacturing process vs. the process that Catalent had initially proposed..  A transfer document was created that retained the critical process parameters, while identifying areas of improvement.  In this project, opportunities were identified in the following processes:

•    Optimize the formulation from a three tank process to a two tank process
•    Incorporate an increased level of automation to the formulation process
•    Redesign of the unit dose BFS ampoule to minimize the probability of leakage
•    Redesign of the label and the label application process
•    Redesign of the foil pouch to reduce scrap experienced at the original site

Per ICH Q8(R), Quality by Design is a systematic approach that begins with predefined objectives and emphasizes product and process understanding and process control, based on sound science and quality risk management. 

Using this definition, the first step of the technical transfer is to create an outline or process flow of the existing process with critical parameters and expected results. 

The outline should also include examination of the ingredients (Active Pharmaceutical Ingredients (API) and excipients) in the formulation to understand any special handling requirements or attributes associated with the materials. 

This product contained two APIs and a polymer solution composed of two grades of the same material.  In the course of the discussion, the customer mentioned a concern over the incorporation or mixing of the polymer solution.  In their experience with this material, the polymer formulation was prone to the formation of globules that could cause inconsistent viscosity results, a critical process parameter, and potential clogging of the clarification filters used in the filling process. 

In addition, in the product manufacturing process, a polymer solution must be bulk sterilized prior to addition of the API solution using sterile filtration.  The high viscosity of the polymer and the uncertainty of the filtration process led the original site to develop a three tank process: a formulation tank for the polymer base with bulk sterilization capability, an API formulation tank, and a third tank to pre-filter the polymer into prior to combining with the API solution.

Both customer and drug service supply partner performed a  risk assessment on the formulation process using a standard approach of assigning values to each step of the process for the rate/likelihood of occurrence, the ability to detect the problem, and the severity of impact to the process in order to properly design a formulation system based on the severity and probability of occurrence.

Catalent engineers believed that the process should dictate the design of equipment, and that a sound process combined with the proper equipment design would be robust enough to overcome any operator variability introduced by personnel performing a task.  For this product, the highest risk to the process was the formulation of the polymer solution.

A solution for the inconsistency in the incorporation of the polymer was presented in the design of a highly automated two-tank formulation system.  The polymer formulation involves the mixing of two lightweight hygroscopic powder ingredients into Water for Injection (WFI) at a high temperature. 

The mixing vessel must have the proper agitation system that creates the proper vortex to incorporate the powder, or the powder will float on the surface of the WFI and form globules.  The Catalent team designed a polymer tank that uses a standard top down mixer and also features a side scrapping system to aid in the incorporation of the powders.  The side scrappers also prevent scorching of the polymer base against the heat transfer surface during the bulk sterilization process. 

In addition, the two-tank process lowered the investment cost for capital equipment by the customer and significantly reduced the necessary floor space in the formulation area.  The design reduced the amount of automation needed, and eliminated the loss of polymer solution inherent in transferring material between tanks.

By automating the formulation process, any variability introduced by the operators was greatly reduced.  Critical parameters such as temperatures, mixing speeds, and mixing times are PLC controlled and monitored, rather than being dependent on operator manipulation.  The sequencing of valves throughout the process is automated and programmed to prevent inadvertent opening or closing of pathways vital to maintaining sterility. 

During the site acceptance testing and engineering tests of the formulation skid, simple Design of Experiments (DOE) were utilized to determine the optimal mixing speeds and durations for each of the polymers. 

The tests were performed using multiple operators to add the powders using techniques varying from a slow addition using a scoop, to rapid addition by pouring the entire volume of powder into the vessel as fast as possible. 

Various shaft lengths and impeller designs were tested prior to selecting a combination that would work with the range of operator induced variability possibly introduced during mixing.  These mixing parameters were then programmed into the automation controls of the skid and incorporated into the Master Batch Record (MBR) for the product.

The results from this design process can be seen in the following graphs from the first eight production batches using this formulation system.  The graphs display the viscosity, pH, osmolality and the assay values for the two APIs in the formulation.  The batches were produced using regular production operators trained on the equipment and the MBR.

The results for API 1 are significant when compared to API 2 (Tables).  Although all the results are acceptable and the Cpk is above 1.00, the difference in performance may indicate the need to improve the consistency of API received, or may be a result of the variability of the analytical method. The storage and dispensing of API 1 will be reviewed to investigate any variation that can be correlated to potency of the final product. This performance will be tracked for API that is sourced from additional suppliers.
This chart of pH results shows a lower process capability and the possible need for a more specific testing method. Combining this capability study with the Gage R&R study proved that we needed to be able to read a higher resolution than we were currently using on our testing methods for pH.These charts show the process capability (Cpk) of the formulation equipment in conjunction with the materials used. Cpk is the index or number which measures how close a process is running to its specification limits, relative to the natural variation of the process. The higher the value of Cpk, the less likely it is that the result will be product outside the product specification limits.  Expressed in Six Sigma terms, a Cpk of 1.33 represents a four sigma level.  The viscosity results from the first eight batches show a Cpk of 8.70 which is a six sigma level of quality.  While requiring 30 samples of data is generally ideal to understand the true capability of a process, this preliminary data is a good indication of the process control that we can expect in the short term and therefore strive to achieve on a consistent basis. All of the formulation test results show a value above 1.00.  If a long term sample process holds a sigma value above 1.00 that is an indicator of the long term capability of the process.The risk assessment also identified three areas in the primary and secondary packaging processes that could be improved.  The customer identified leaking units and excess plastic from the de-flashing process following the formation and filling of the ampoules as problematic at their current manufacturing site.  Catalent engineers proposed an improvement in the individual vial or pipette design that would improve the container closure integrity and solve the excess material problem.  The original vial required de-flashing of the area between the neck of each vial as well as separation of each five unit card after the units exit the BFS machine.  While maintaining the same tab and body foot print, Catalent improved the primary container by reinforcing the neck area and eliminated the removal of plastic between the individual pipettes.  If excess plastic remained with the de-flashing of the original vial, they became rejects.  If they were not caught in the inspection process, the downstream secondary packaging process was impacted in labeling and pouching.  The Catalent design eliminated the problems caused by the additional de-flashing which improved the overall performance of the product during packaging.Improvements were also proposed by Catalent to the secondary packaging when discussions with the customer revealed an excessive reject rate during labeling.  The original process uses a single wrap around label that covers both sides of the tabs on each of the five vials.  This label application system generated a re-processing rate of approximately 5% due to label alignment.  To reduce the re-processing rate, Catalent proposed separate labels for each side of the tab.  The specifications for the labeler designed to place labels on both sides needed to deliver the precision necessary to properly align the groups of five perforated labels in the required space with little margin for error. With Catalent’s experience in providing packaging solutions, and our understanding of the capabilities of our vendors and machine suppliers this was accomplished.  There-processing rate for label defects (missing or poor alignment) has been reduced to less than 1%.A final Catalent recommendation for improvement involved replacing the type of foil used in the over pouch at the original manufacturing site with a standard foil used at the Woodstock site.  The foil pouch serves as a moisture barrier and provides protection from light.  The Catalent foil selected would meet these storage and stability requirements and would eliminate the need to acquire specialized equipment to accommodate the original foil.An added benefit to the change would be reduction of the re-processing experienced at the original site which occurred at a rate of 20% on average.  The original foil was a very rigid two- layer paper-based material that could only be run on a custom built machine.  The original pouch is heat sealed and crimped on all four sides.  A set of three BFS cards is placed vertically into the foil and any minor misalignment of the cards will cause a crimping reject.  The Catalent design uses a two-layer, lighter weight, aluminum foil that has a bottom fin seal and sealing at both ends.  The cards are presented horizontally into the pouch machine using well defined guides to maintain the alignment and stability needed to reduce defects.  The re-processing rate has been decreased to less than 2% for foil pouches.  Of course, this change could not be implemented until stability testing was performed to show that the packaging change had no product impact.Combining regulatory approval for the tech transfer along with process capability and other continuous improvements is an efficient way to introduce improvements.   In this project, all changes proposed were approved during the transfer process, resulting in a superior process to consistently produce quality product.   It is important to start with a process comparison, utilize risk assessment and data driven design tools for proposed changes and to include those documents in the filing.  Using the principles of Quality by Design should be part of the technical transfer process.  A manufacturing partner should provide data driven solutions that result in better processes and a reliably supplied product. About the AuthorsNorman Weichbrodt is Strategic Account Manager and Mike Camalo is Six Sigma Black BeltAt Catalent Pharma Solutions. Woodstock, IL
About the Author

Norman Weichbrodt and Mike Camalo | Catalent Pharma Solutions